Hydrogen Bond Variations of Influenza A Viruses During Adaptation in Human

Luo, Jiejian; Deng, Lizong; Ding, Xiao; Quan, Lijun; Wu, Aiping; Jiang, Taijiao

doi:10.1038/s41598-017-14533-3

Download PDF

Article
Open access
Published: 30 October 2017

Hydrogen Bond Variations of Influenza A Viruses During Adaptation in Human

Jiejian Luo^1,2^na1,
Lizong Deng^3,4^na1,
Xiao Ding^3,4,
Lijun Quan^3,4,
Aiping Wu^3,4 &
…
Taijiao Jiang^1,3,4

Scientific Reports volume 7, Article number: 14295 (2017) Cite this article

934 Accesses
3 Citations
Metrics details

Subjects

Abstract

Many host specific mutations have been detected in influenza A viruses (IAVs). However, their effects on hydrogen bond (H-bond) variations have rarely been investigated. In this study, 60 host specific sites were identified in the internal proteins of avian and human IAVs, 27 of which contained mutations with effects on H-bonds. Besides, 30 group specific sites were detected in HA and NA. Twenty-six of 36 mutations existing at these group specific sites caused H-bond loss or formation in at least one subtype. The number of mutations in isolations of 2009 pandemic H1N1, human-infecting H5N1 and H7N9 varied. The combinations of mutations and H-bond changes in these three subtypes of IAVs were also different. In addition, the mutations in isolations of H5N1 distributed more scattered than those in 2009 pandemic H1N1 and H7N9. Eight wave specific mutations in isolations of the fifth H7N9 wave were also identified. Three of them, R140K in HA, Y170H in NA, and R340K in PB2, were capable of resulting in H-bond loss. As mentioned above, these host or group or wave specific H-bond variations provide us with a new field of vision for understanding the changes of structural features in the human adaptation of IAVs.

Evolutionary, genetic, structural characterization and its functional implications for the influenza A (H1N1) infection outbreak in India from 2009 to 2017

Article Open access 11 October 2019

A risk marker of tribasic hemagglutinin cleavage site in influenza A (H9N2) virus

Article Open access 15 January 2021

Characterizing genetic and antigenic divergence from vaccine strain of influenza A and B viruses circulating in Thailand, 2017–2020

Article Open access 12 January 2021

Introduction

Influenza A viruses (IAVs) are negative-sense, single-stranded, and segmented RNA viruses, whose natural reservoir is wild aquatic bird. Currently, H1N1 and H3N2 IAVs co-circulate amongst human worldwide seasonally, which cause more than 5 million cases of severe illness and about 500 thousand deaths every year¹. In theory, avian IAVs are not capable of infecting human because of the host-range restriction². However, the emergence of human infections with avian H5N1 and H7N9 IAVs in these years demonstrates a potential pandemic threat^3,4. Unfortunately, it is still unclear how IAVs adapted in different hosts. Previous researches have found that the HA protein plays a crucial role in the host adaptation because it binds to sialic acid receptors of host cells and mediates membrane fusion and viral entry². In general, the HA proteins from human-adapted IAVs tend to bind a2,6-linked sialic acid linkages while those from avian-adapted IAVs prefer a2,3-linked sialic acid linkages⁵. In addition, other viral proteins, such as polymerase subunits, have also been reported as a determinant of host range of IAVs^2,6.

In the last decade, computational or experimental researches have been carried out to identify singular or combinatorial host specific signatures of IAVs^7,8,9,10, some of which were likely to facilitate the host adaptation process. However, the analysis of mutations generally focused on amino acid changes instead of structural variations of proteins. The hydrogen bond (H-bond) is one of the most important noncovalent interactions in biology which plays a significant role in stabilizing the three-dimensional structures and molecular interactions¹¹. Previous studies have identified several mutations with H-bond variations in the process of host adaptation of IAVs. Xu et al. showed that the dual mutations E190D and G225D at the HA receptor binding sites switched the receptor specificity from avian-type to human-type in 2009 pandemic H1N1 (pH1N1) because of the formation of H-bond interactions between the glycan and HA¹². In addition, the mutation H110Y which is located at the trimer interface forms a H-bond with the 413N of the adjacent monomer in order to stabilize the trimeric HA protein of H5 subtype¹³. The NA of H5N1 and pH1N1 with H274Y mutation significantly weakened the binding affinity for the anti-viral drug oseltamivir, which resulted from the loss of H-bond interactions between the oseltamivir and two residues of NA (178W and 152R)¹⁴. Moreover, co-mutations V344M and I354L in the PB2 subunit of pH1N1 enhanced binding affinity by creating additional H-bond contacts between PB2 cap binding domain and the pre-mRNA cap analogue m7GTP¹⁵. However, these researches, as stated, were specific to a few influenza subtypes and only covered a few of proteins. Here, the H-bond variations of host specific and group specific sites in viral proteins were systematically evaluated. The combinations of mutations and H-bond changes at these sites significantly varied among pH1N1, human-infecting H5N1 and H7N9. In addition, the wave specific sites of the fifth H7N9 wave and their corresponding effects on H-bonds were also investigated.

Results

The H-bond variations of host specific sites in the eight internal proteins

We assessed the H-bond variations of viral internal proteins between avian and human IAVs. As shown in Table S3, a total of 36999 non-redundant internal protein sequences (M1: 1635, M2: 2184, NP: 4482, NS1: 4479, NS2:1939, PA: 7733, PB1: 6973, PB2: 7574) were included in avian dataset. For human dataset, it contained two seasonal subtypes H1N1 and H3N2. There were 2781 sequences (M1: 165, M2: 253, NP: 294, NS1: 441, NS2: 171, PA: 411, PB1: 512, PB2: 534) in H1N1 and 14457 sequences (M1: 537, M2: 919, NP: 1442, NS1: 2478, NS2: 534, PA: 2643, PB1: 2737, PB2: 3167) in H3N2.

Sixty host specific sites were identified in the eight internal proteins (M1, M2, NS1, NS2, NP, PA, PB1, and PB2) of avian and human IAVs (Table 1). Over half of them (32/60) were in the viral RNA polymerase, including 13 sites in NP, 3 sites in M1, 6 sites in M2, and 6 sites in NS1. As host-associated positions reported in previous literatures^9,10, the left two sites 70 and 107 of NS2 with dScore (0.89 and 0.88, respectively) below the threshold of 0.90 were excluded from our study. The H-bond variations of these host specific sites were evaluated through the differences of H-bond contacts with their neighboring residues between before and after a mutation. The relative solvent accessibility (RSA) of all the sites was calculated. Sites with RSA value above 25% were defined as exposed sites (located on the surface of protein). And then they were mapped onto the linear sequences of the proteins with functional annotations (Fig. 1). As shown in Fig. 1, 27 host specific sites contained H-bond loss or formation causing mutations including 10 sites in PA, 10 sites in PB2, 4 sites in NP, 2 sites in M1, and 1 site in NS1. There is no significant difference between the distributions of the H-bond variation sites and non H-bond variations sites on the three dimensional structures (the ratio of exposed sites: 77.8% in H-bond variation sites VS 63.6% in non H-bond variation sites, the two-tailed Fisher’s exact p-value is 0.27). As shown in Table 2, the number of sites at which mutations only gave rise to one kind of effect on H-bonds was 13 for H-bond loss and 10 for H-bond formation. The mutations at the other four sites could result in both H-bond loss and formation. For the convenience, all the H-bonds were written in the format of ‘HB(donor residue, acceptor residue, donor atom–H…acceptor atom)’. Notably, different mutations at the same site would lead to similar H-bond variations. Both S421I and S421V in the C-terminal domain of the PA protein could disrupted the H-bond HB(490R, 421S, Nη2–H…Oγ). Interestingly, 421I and 421V of PA were the dominant residue in H1N1 and H3N2, respectively. The same phenomena existed at NP 375, PA 65, and PB2 81 (Table 2).

Table 1 60 host specific sites of internal proteins in avian and human IAVs.

Full size table

Table 2 H-bond variations at host specific sites of internal proteins.

Full size table

The H-bond variations of group specific sites for the HA and NA proteins

To analyze the H-bond variations in HA and NA, we also collected and selected HA and NA subtypes with more than 100 non-redundant sequences (Tables S1 and S4). For human dataset, 1560 H1, 10837 H3, 1393 N1, and 9512 N2 non-redundant protein sequences were included. For avian dataset, there were a total of 11284 HA proteins (H1: 139, H2: 295, H3: 819, H4: 820, H5: 3617, H6: 1080, H7: 1153, H8: 116, H9: 1951, H10: 489, H11: 498, H12: 159, H13: 148) and 9201 NA proteins (N1: 2321, N2: 2546, N3: 714, N4: 141, N5: 221, N6: 1239, N7: 459, N8: 1054, N9: 506).

We were unable to detect any universal host specific site among all subtypes of the HA/NA protein. Then, the mutation analyses of the HA and NA proteins were done at the group level. Nevertheless, few sites could be identified when all the ten subtypes of the group 1 HA were considered (Table S1). To capture enough differential signatures in group 1 HA, we just selected H1, H2, H5, and H6 subtypes (a sub-group of group 1 HA) for calculation. The number of detected group specific sites detected in group 1 HA (H1, H2, H5, and H6 subtypes considered), group 2 HA (H3, H4, H7, and H10), group 1 NA (N1, N4, N5, and N8) and group 2 NA (N2, N3, N6, N7, and N9) was 8, 7, 9, and 6, respectively (Table 3). Notably, although sites 190 and 225 were overlapped in two groups of the HA protein, their amino acid usages in human infections were slightly different. For the HA protein of human infections (H1 lineage in group 1 and H3 lineage in group 2), the dominant residues of site 190 and 225 were both Asp. However, a certain proportion of human H1 possessed Asn at position 190 while some of human H3 had Asn at position 225 (Table 3). The H-bond variations related sites on protein structures between two groups for both HA and NA proteins were significantly different (Fig. 2a and b).

Table 3 Group specific sites of HA and NA.

Full size table

As shown in Fig. 2c–f, twenty-six of 36 mutations at the group specific sites caused H-bond loss or formation in at least one subtype, 17 of which didn’t share the same H-bond changes in all subtypes of the same group. For example, the mutation Q226I of group 2 HA led to H-bond loss in H3 and H7 subtypes, whereas it didn’t give rise to H-bond changes in H4 and H10 subtypes (Fig. 2d). In addition, the loss of H-bond contacts in H3 was HB(226Q, 98Y, Nε2-H…Oη) while those in H7 were HB(136T, 226Q, Oγ1-H…Oε1) and HB(226Q, 137T, Nε2-H…Oγ1) (Table S2). The residue at site 226 in the receptor binding pocket (RBP) of the HA protein was critical for receptor specificity for an avian or mammalian host¹⁶. We constructed the RBP superposition model and found that the local structures at position 226 in group 2 HA were different to some extend (Fig. 2g).

H-bond variations in pandemic and sporadic human-infecting IAVs

The comparison of the H-bond variations between pandemic and sporadic human-infecting IAVs was performed in Fig. 3a. Among the four pandemic representative isolations, A/Albany/20/1957 (H2N2, 1957) and A/Aichi/2/1968 (H3N2, Aichi2) contained more human-preferential mutations at the 27 H-bond variation sites than A/Brevig Mission/1/1918 (H1N1, 1918) and A/California/04/2009 (H1N1, CA04). For the five representative human-infecting IAVs, mutations were sporadic and their combinations were significantly different from the four pandemic strains.

The reassortant pH1N1 has co-circulated with H3N2 seasonally since 2009, but it is different from the seasonal H1N1 before 2009¹⁷. In the meantime, H5N1 and H7N9 IAVs are the two major subtypes of avian IAVs that can cause large-scale infections in human and poultry¹⁸. As reported in previous researches, the pattern of the spread of H5N1 in humans and birds around the world is consistent with the wild bird migration and poultry trade activities. In contrast, human cases of H7N9 and isolations of H7N9 in birds and the environment have largely occurred in a number of contiguous provinces in south-eastern China¹⁸. In addition, it has been found that the H7N9 cases are mainly among older cohorts while H5N1 cases are among younger cohorts¹⁹. Thus, it was necessary to do further comparisons of these three subtypes of IAVs to investigate their adaptations to human.

The number of the host specific or group specific sites with mutations in isolations of pH1N1, human-infecting H5N1 and H7N9 was 20, 11, and 10, respectively (Fig. 3b–d). Besides, the combination patterns of mutations and H-bond changes in these three subtypes of IAVs were also different. There were more mutations in the NP protein of pH1N1 than those of H5N1 and H7N9. As shown in Fig. 3b, there were 8 mutations that caused H-bond loss or formation in all seasons of pH1N1 except NA S372K and PA N321K. The amino acids at both NA 372 and PA 321 were Asn in isolations of season 09–10, which were replaced by Lys from season 10–11 on. It was obvious that the mutations of H5N1 distributed more scattered than those of pH1N1 and H7N9. The H5N1 infections in human were divided into two emergences. The first emergence was in Hong Kong in 1997 and the re-emergence was in Mainland China in 2003¹⁸. Two mutation patterns in H5N1 matched with these two emergences (Fig. 3c). There were five mutations with H-bond variations in most of the H7N9 isolations, among which HA Q226L/I, HA E479A, and PB2 E627K caused H-bond loss while NS1 A60E and PA A337T resulted in H-bond formation (Fig. 3d). The patterns of H-bond variations between H5N1 and H7N9 were significantly different (Fig. 3c and d). The NS1 A60E and PB2 E627K were the two common mutations in both H5N1 and H7N9 viruses. The NS1 A60E causing the H-bond formation was mainly existed in H5N1 strains collected in 1997 and H7N9 strains while H5N1 strains collected after 2003 preferred amino acid Ala at NS1 60. Less H5N1 isolations possessed the E627K mutation than the H7N9 strains. In addition, the other mutations that caused H-bond loss or formation in H5N1 were more sporadic than those in H7N9.

Wave specific mutations and H-bond variations in the fifth H7N9 wave

The H7N9 virus has caused five epidemic waves of human infections in China since its first emergence in 2013. It was noted that elevated morbidity and mortality in a wider affected area were observed in the fifth wave²⁰. To investigate the extraordinary phenomenon for the fifth wave, mutation comparison was also performed between the H7N9 isolations derived from wave 5 and from wave 1–3 (sequences data of wave 4 were not enough). As shown in Table 4, 8 mutations specific to the fifth wave were identified, which contained 5 mutations in the HA protein, 1 mutation in the NA protein and 2 mutations in the PB2 protein. The mutations R140K in the HA protein, Y170H in the NA protein and R340K in the PB2 protein were capable of causing the H-bond loss (Fig. 4a). The location of the residue 140R was in the 123–149 loop which was near the conserved B-cell epitopes 123–134 region (MGFTYSGIRTNG) of avian H7 HA. The mutation from Arg to Lys at position 140 disrupted the H-bond interaction HB(140R, 141R, Nη1-H…O) (Fig. 4b and c). The mutation Y170H in the NA protein caused the loss of a H-bond between the side-chains of Tyr-170 and Asp-113 which was in 163–172 loop (LSSPPTVYNS) and 111–120 loop (SSDVLVTREP) respectively (Fig. 4d and e). Besides, these two residues were both located on the interface of two subunits of NA (Figure S2). In addition, the R340K mutation in the PB2 protein could break the two H-bond interactions with its neighbor residues 358 and 342 (Fig. 4f and g). It is known that the Lys-340 in the cap bind pocket of PB2 played critical roles in mammalian adaptation of the H10N8 virus and viruses harboring PB2-588V exhibited higher polymerase activity²¹. As it happened, a majority of the H7N9 strains in the fifth wave contained both 340 K and 588 V in PB2 (Fig. 4a).

Table 4 Wave specific mutations in the fifth H7N9 wave.

Full size table

To investigate the difference between the Yangtze River Delta lineage (YRD) and the Pearl River Delta lineage (PRD), we collected the strains of H7N9 from wave 3 to wave 5 and constructed the approximately-maximum-likelihood phylogenetic tree of HA protein (Figure S6). The PRD and YRD lineages of the fifth H7N9 wave were colored as yellow and blue bar in Figure S6, respectively. Finally, 9 PRD and 114 YRD strains with completed genomes in the fifth H7N9 wave were selected.

As shown in Table S8, differential sites and H-bond variations between the PRD and YRD lineages in the fifth H7N9 wave were evaluated. The H-bond variations were assessed with the A/Anhui/1/2013 as reference. In total, 19 differential regions or sites were found, including 8 sites in HA, 1 site in M1, 1 site in M2, 8 sites in NA, and 1 site in PA. The insertion of basic amino acid residues RKRT at the cleavage site connecting the HA1 and HA2 peptide region was found in all the 9 PRD strains of wave 5, which was a signature of highly pathogenic avian influenza viruses²². In addition, the number of H-bond variations sites between PRD and YRD of wave 5 was 9, including 3 sites in HA, 5 sites in NA, and 1 site in PA (Table S8). For the lack of sufficient PRD sequences of wave 5 H7N9, the differences or signatures between PRD and YRD linages found here needed to be further validated.

Discussion

Typical avian IAVs don’t have the capacity to replicate efficiently and cause human infections. In order to become capable of establishing in human, avian IAVs must overcome species barriers and adapt to a new host environment. Some changes need to be done to maintain the stability or function of viral proteins during the human adaptation of IAVs². The H-bond is one of the most important noncovalent interactions for protein stabilization and molecular interactions¹¹. Investigation of the changes of H-bond features will promote understanding the mechanism of viral adaptation in human.

In our analysis, 60 host specific sites of internal proteins between avian and human IAVs were identified, 27 of which contained mutations with effects on H-bonds. 75% (45/60) of the host specific sites were in the RNA polymerase and NP proteins. The RNA polymerase is responsible for the transcription and replication of the virus genome, while the NP encapsulates the virus genome to form a ribonucleoprotein (RNP) particle for the purposes of transcription and packaging²³. It is well documented that polymerases from avian IAVs don’t function well in mammalian host²⁴. This high proportion of mutations in the RNP might play important roles in viral adaptation in human. H3N2 and H1N1 are two major lineages of human IAVs. Despite their common origin, the internal protein sets of these two lineages have evolved independently²⁵. Some sites with different residue usages in H1N1 and H3N2, such as PA 421 and NP 375, had the same effects on H-bonds (Table 2), which suggests the diversity of human adaptation. Group specific sites were further identified in HA and NA, which were shown in Table 3. The H-bond variations of some mutations at group specific sites of the HA/NA proteins were different among different subtypes (Fig. 2c–f). On the one hand, it might result from the differential local structures. On the other hand, these mutations had other important functions we haven’t yet discovered besides the H-bond contacts in the protein. Although the mutation Q226I was identified in H3, H4, H7, and H10 subtypes, the H-bond variation only emerged in the H3 and H7 subtypes. The local structural difference of 226Q was clear in the RBP superposition model of H3, H4, H7, and H10 subtypes (Fig. 2g). It has been well proved that Q226I/L in the RBP increased the ability to bind a2–6 sialic acid^{26,27,28,29,30}, which implies its role in inter-molecular interactions is more important than that in intra-molecular interactions in HA. Besides, dual mutations E190D and G225D of receptor binding sites had no effect on H-bonds in H1 HA according to our assessment (Fig. 2c), but they mediated several H-bonds interactions between the glycan and HA to switch the receptor specificity from avian to human in H1N1 subtypes¹². In other words, mutations with no effect on H-bonds in viral proteins might be an important part of inter-molecular interactions. Unfortunately, we were unable to evaluate these inter-molecular interactions systematically due to the insufficient structural information of protein or molecular interaction.

It is worth noting that we couldn’t give an evaluation of H-bond variations due to co-mutations or multiple mutations in this study. The mutation E627K in the PB2 627 nuclear localization signal (NLS) domain of the RNA polymerase disrupted the sidechain-sidechain H-bond interaction with 591Q (Table 2). Actually, it’s known that PB2 E627K can alter the surface electrostatic potentials of PB2-627NLS domain with the assistance of residues at 590 and 591³¹. According to our statistics, the major amino acids at 590 and 591 of PB2 in both avian and human IAVs were Gly and Gln, respectively. In the meantime, we assessed the effect of the mutation on H-bonds based on the avian consensus sequence environment without consideration of residue combinations in the real strains. Besides, the crystal structure of RNA polymerase of IAVs used in homology modeling was bat origin (A/little yellow-shouldered bat/Guatemala/060/2010, H17N10, PDB ID :4WSB) and its chain sequences were greatly different from the ones in other subtypes, which might result in structural differences. Therefore, more accurate and suitable crystal structures are needed to validate these H-bond variations resulting from adaptive mutations.

Both H5N1 and H7N9 have caused sporadic human cases without any evidence of sustained and human-to-human spread, but their patterns of H-bond variations were significantly different (Fig. 3c and d). In fact, the pattern of the spread of H5N1 in humans and birds around the world is consistent with the wild bird migration and poultry trade activities. In contrast, human cases of H7N9 and isolations of H7N9 in birds and the environment have largely occurred in a number of contiguous provinces in south-eastern China¹⁸. Besides, it has been found that the H7N9 cases are mainly among older cohorts while H5N1 cases are among younger cohorts, and the lifelong protection against H5N1 and H7N9 is via different childhood hemagglutinin imprinting¹⁹. So the different epidemic patterns of IAVs and different human immune responding to IAVs may be the possible explanations of differential patterns of mutations and H-bond variations between H5N1 and H7N9.

The H7N9 virus has caused five waves of human infections in China since March 2013. An increased pathogenicity in a wider affected area was observed in the fifth wave²⁰. There is a sufficient preponderance of observed mutations in isolates of the fifth wave when compared with those in wave1–3 at the eight characteristic sites (Table 4). The mutations were acquired in the several strains of wave 4 (Fig. 4a), but we were not sure whether their frequencies were similar to those in wave 5 due to lack of enough sequences. The dual mutations, R340K and A588V in PB2, appeared in most of the isolations of wave 5 (Fig. 4a). The substitution from Arg to Lys at position 340 could disrupt the H-bond interactions with 358E and 342E in the cap binding domain of PB2 (Fig. 4f and g). In fact, both Arg and Lys were basic amino acids with similar chemical properties, whereas their differences of side-chain conformations at PB2 340 were clear. Because all the H-bond calculations were based on homology modeling using CISRR in our study, these H-bond variations need further validation in accurate crystal structures. The mutation from Lys to Asn reduced polymerase activity of A/Hamburg/NY1580/09 strain³². The residue 588V, located in the PB2 627-domain near the polymorphic 590 and 591 residues, had been reported that it is important for H7N9 and H10N8 virus replication and virulence²¹. The dual mutations R340K and A588V in PB2 might be a feature of the fifth wave of H7N9.

In summary, our study gave a systematic assessment of intra-molecular H-bond interactions at host specific or group specific sites between avian and human IAVs, which is helpful for us to understand human adaptation of IAVs from a new perspective. Of course, the H-bond interaction is just one kind of the noncovalent interactions. The effect of mutations may be multi-functional and they tend to function together. Therefore, more efforts need to be put into the study of the variations of structural features to get a comprehensive understanding of how these mutations work.

Material and Methods

Datasets

We retrieved all full-length sequences for ten proteins (HA, NA, NS1, NS2, M1, M2, NP, PA, PB1, and PB2) of IAVs isolations between 1918 and June 2017 from the GISAID database (http://platform.gisaid.org/epi3/frontend) and the Influenza Virus Database in NCBI³³. Sequences from these two databases were merged. For human IAVs, we mainly considered epidemic seasonal H1N1, H2N2 and H3N2 strains. The human H2N2 contained less than 50 non-redundant sequences in each protein (Table S3) and this subtype was not considered in our statistical analysis. As reported in previous researches, the internal protein of human H1N1 and H3N2 have evolved independently²⁵. It’s necessary to compare these two subtypes separately. Our investigation was focused on H-bond variations between avian and human IAVs and how these features changed after avian IAVs overcome host barriers to establish sustained infections in human. The effect of IAV reassortments should be excluded. Thus, the pandemic strains (1918 H1N1pdm, 1957–1958 H2N2pdm, 1968 H3N2pdm, and 2009 H1N1pdm) were removed from human dataset as they were reassortants^17,34,35,36. For avian IAVs, all subtypes of IAVs were considered except suspicious subtype H1N1, H2N2, and H3N2 which also circulated amongst human. Besides, those strains annotated as mixed subtypes or lab strains were also excluded. Moreover, we did additional sequence cleaning for NS segment. NS gene could be grouped into two major variants known as allele A and B and human NS basically belonged to allele A^37,38. So it was reasonable to remove allele B sequences before comparison of NS segment between avian and human. Sequence identities between allele A and B were about 70%, whereas those within each allele were above 90%³⁹. We eliminated allele B sequences according to identities, in which case the NS of A/tern/South Africa/1961 (allele A; accession: CY014988) and A/redhead duck/ALB/74/1977 (allele B; accession: CY004739) were chosen as references. Reassortants at a given period such as 2009 pandemic H1N1 appear to be highly similar. We collapsed identical sequences for each protein with identity threshold equal to 1 using cd-hit⁴⁰. This step is necessary and important for reduce the proportion of potential unknown reassortants in our dataset, although the frequency of IAV reassortments was low and few inter-subtype reassortants have actually established sustained infections in human^41,42. Finally, we got a non-redundant dataset that comprised human IAVs (human-host H1N1 and H3N2) and avian IAVs (avian-host subtypes excluding H1N1, H2N2 and H3N2) (Table S3).

It is widely known that the HA protein can be divided into two groups: group 1 HA (H1, H2, H5, H6, H8, H9, H11, H12, H13, H16, H17, H18) and group 2 HA(H3, H4, H7, H10, H14, H15)⁴³. The NA protein also has two groups: group 1 NA (N1, N4, N5, N8) and group 2 NA (N2, N3, N6, N7, N9)⁴⁴. The two groups of HA can be further divided into several subgroups (Tables S1 and S4). The H14, H15, H17 and H18 subtypes with few sequences were not considered in our analysis.

To validate the reasonability of our datasets, we constructed the approximately-maximum-likelihood phylogenetic trees of each protein with FastTree 2.1 (http://www.microbesonline.org/fasttree/). For each of the internal proteins, three sub-clades could be achieved: avian clade, human H1N1 clade, and human H3N2 clade (Figure S3). For the HA or NA protein, two groups were achieved in either avian or human dataset (Figures S4 and S5).

Sequences of each internal protein were aligned by MAFFT version 7⁴⁵. Because of low sequence similarities between subtypes of HA(NA), a structure based sequence alignment should be constructed, in which case sequences of HA (NA) were added into using MAFFT with ‘–add’ parameter. Structure based sequence alignment for HA described in literature⁴⁶ was used. As for NA, crystal structures of N1 to N9 were downloaded from PDB database⁴⁷ and aligned with structure alignment tool DeepAlign⁴⁸.

Identification of specific sites

Given a column of two aligned sequence sets (set A and set B), the frequencies of residues of the column in each set were counted. The dScore was defined to assess the difference of a certain site between two sets with the following formula subsequently:

$${\rm{dscore}}\,({\rm{c}})=1-\,\sum _{r\in R}\,{\rm{\min }}({f}_{A}(r,c),{f}_{B}(r,c))$$

(1)

r is an arbitrary residue in the standard amino acids set R. f_A(r,c) is denoted as the frequency of residue r in column c of in set A while f_B(r,c) is the frequency of residue r in column c of in set B. dScore(c) ranges from 0 to 1.The more the dScore(c) approximates to 1, the greater different the site c between two sets is.

To balance the sequences between different lineages or subtypes, we used a bootstrap sample method. First of all, 500 sequences of each lineage or subtype were sampled with replication and performed one calculation of dScore with equation (1). This procedure was repeated 1000 times and an average dScore was obtained. Finally, sites with average dScore more than 0.90 were selected (Figure S1).

Homology modeling

We predicted the protein structures using the side-chain modeling tool CISRR⁴⁹. Crystal structures with high resolutions (<3.0Å) were selected as templates in priority (Table S5).

Identification of H-Bonds

H-bonds were identified using the simple geometric criteria of Baker and Hubbard¹¹.The distance between donor atom and acceptor atom⁵⁰ is less than 3.5 Å and the angle between the donor antecedent, donor and acceptor 90–180°. Main-chain and main-chain H-bonds were not considered in our analysis.

Calculation of Relative Solvent Accessibility

Relative solvent accessibility (RSA) of a residue was calculated using the program NAccess (unpublished, S. Hubbard and J. Thornton 1992–6, http://www.wolf.bms.umist.ac.uk/naccess/) and ACCpro⁵¹. A site was regarded as exposed if its RSA was above 25%⁵².

References

Lozano, R. et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: A systematic analysis for the Global Burden of Disease Study 2010. Lancet 380, 2095–2128 (2012).
Article PubMed Google Scholar
Cauldwell, A. V. et al. Viral determinants of influenza A virus host range. J. Gen. Virol. 95, 1193–1210 (2014).
Article CAS PubMed Google Scholar
De Jong, M. D. & Hien, T. T. Avian influenza A (H5N1). J. Clin. Virol. 35, 2–13 (2006).
Article PubMed Google Scholar
Watanabe, T., Watanabe, S., Maher, E. A., Neumann, G. & Kawaoka, Y. Pandemic potential of avian influenza A (H7N9) viruses. Trends Microbiol. 22, 623–631 (2014).
Article CAS PubMed PubMed Central Google Scholar
Stevens, J. et al. Glycan microarray analysis of the hemagglutinins from modern and pandemic influenza viruses reveals different receptor specificities. J. Mol. Biol. 355, 1143–1155 (2006).
Article CAS PubMed Google Scholar
Manz, B. et al. Adaptation of Avian Influenza A Virus Polymerase in Mammals To Overcome the Host Species Barrier. J. Virol. 87, 7200–7209 (2013).
Article PubMed PubMed Central CAS Google Scholar
Tamuri, A. U., Dos Reis, M., Hay, A. J. & Goldstein, R. A. Identifying changes in selective constraints: Host shifts in influenza. PLoS Comput. Biol. 5 (2009).
Finkelstein, D. B. et al. Persistent host markers in pandemic and H5N1 influenza viruses. J. Virol. 81, 10292–10299 (2007).
Article CAS PubMed PubMed Central Google Scholar
Chen, G. W. et al. Genomic signatures of human versus avian influenza A viruses. Emerg. Infect. Dis. 12, 1353–1360 (2006).
Article CAS PubMed PubMed Central Google Scholar
Miotto, O. et al. Complete-proteome mapping of human influenza A adaptive mutations: Implications for human transmissibility of zoonotic strains. PLoS One 5 (2010).
Sticke, D. F., Presta, L. G., Dill, K. A. & Rose, G. D. Hydrogen bonding in globular proteins. J. Mol. Biol. 226, 1143–1159 (1992).
Article Google Scholar
Xu, R., McBride, R., Nycholat, C. M., Paulson, J. C. & Wilson, I. A. Structural Characterization of the Hemagglutinin Receptor Specificity from the 2009 H1N1 Influenza Pandemic. J. Virol. 86, 982–990 (2012).
Article CAS PubMed PubMed Central Google Scholar
Zhang, W. et al. An Airborne Transmissible Avian Influenza H5 Hemagglutinin Seen at the Atomic Level. Science (80-.). 340, 1463–1467 (2013).
Article ADS CAS Google Scholar
Malaisree, M. et al. Source of oseltamivir resistance in avian influenza H5N1 virus with the H274Y mutation. Amino Acids 37, 725–732 (2009).
Article CAS PubMed Google Scholar
Bhoye, D., Behera, A. K. & Cherian, S. S. A molecular modelling approach to understand the effect of co-evolutionary mutations (V344M, I354L) identified in the PB2 subunit of influenza A 2009 pandemic H1N1 virus on m7GTP ligand binding. J. Gen. Virol. 97, 1785–1796 (2016).
Article CAS PubMed Google Scholar
de Vries, R. P. et al. Hemagglutinin Receptor Specificity and Structural Analyses of Respiratory Droplet-Transmissible H5N1 Viruses. J. Virol. 88, 768–773 (2014).
Article PubMed PubMed Central CAS Google Scholar
Michaelis, M., Doerr, H. W. & Cinatl, J. An influenza A H1N1 virus revival - Pandemic H1N1/09 virus. Infection 37, 381–389 (2009).
Article CAS PubMed Google Scholar
Bui, C. et al. A Systematic Review of the Comparative Epidemiology of Avian and Human Influenza A H5N1 and H7N9:Lessons and Unanswered Questions. Transbound. Emerg. Dis. 63, 602–620 (2016).
Article CAS PubMed Google Scholar
Gostic, K. M., Ambrose, M., Worobey, M. & Lloyd-Smith, J. O. Potent protection against H5N1 and H7N9 influenza via childhood hemagglutinin imprinting. Science (80-.). 354, 722–726 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Huo, X. et al. Significantly elevated number of human infections with H7N9 virus in Jiangsu in eastern China, October 2016 to January 2017. Euro Surveill. 22, 1–10 (2017).
Article Google Scholar
Xiao, C. et al. PB2-588 V promotes the mammalian adaptation of H10N8, H7N9 and H9N2 avian influenza viruses. Sci. Rep. 6, 19474 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Su, S. et al. Epidemiology, Evolution, and Pathogenesis of H7N9 Influenza Viruses in Five Epidemic Waves since 2013 in China. Trends Microbiol. xx, 1–16 (2017).
Google Scholar
Gabriel, G. & Fodor, E. Molecular determinants of pathogenicity in the polymerase complex. Curr. Top. Microbiol. Immunol. 385, 35–60 (2014).
PubMed Google Scholar
Naffakh, N., Tomoiu, A., Rameix-Welti, M. A. & van der Werf, S. Host restriction of avian influenza viruses at the level of the ribonucleoproteins. Annu Rev Microbiol 62, 403–424 (2008).
Article CAS PubMed Google Scholar
Kendal, A. P., Noble, G. R., Skehel, J. J. & Dowdle, W. R. Antigenic similarity of influenza A(H1N1) viruses from epidemics in 1977-1978 to ‘Scandinavian’ strains isolated in epidemics of 1950-1951. Virology 89, 632–636 (1978).
Article CAS PubMed Google Scholar
Hanson, A. et al. Identification of Stabilizing Mutations in an H5 Hemagglutinin Influenza Virus Protein. J. Virol. 90, 2981–2992 (2016).
Article CAS PubMed Central Google Scholar
Imai, M. et al. Experimental adaptation of an influenza H5 HA confers respiratory droplet transmission to a reassortant H5 HA/H1N1 virus in ferrets. Nature, https://doi.org/10.1038/nature10831 (2012).
Matrosovich, M. et al. Early Alterations of the Receptor-Binding Properties of H1, H2, and H3 Avian Influenza Virus Hemagglutinins after Their Introduction into Mammals. J. Virol. 74, 8502–8512 (2000).
Article CAS PubMed PubMed Central Google Scholar
Chen, L.-M. et al. In vitro evolution of H5N1 avian influenza virus toward human-type receptor specificity. Virology 422, 105–113 (2012).
Article CAS PubMed Google Scholar
Herfst, S. et al. Airborne Transmission of Influenza A/H5N1 Virus Between Ferrets. Science (80-.). 336, 1534–1541 (2012).
Article ADS CAS PubMed Central Google Scholar
Kuzuhara, T. et al. Structural basis of the influenza A virus RNA polymerase PB2 RNA-binding domain containing the pathogenicity-determinant lysine 627 residue. J. Biol. Chem. 284, 6855–6860 (2009).
Article CAS PubMed PubMed Central Google Scholar
Otte, A. et al. Adaptive Mutations That Occurred during Circulation in Humans of H1N1 Influenza Virus in the 2009 Pandemic Enhance Virulence in Mice. J. Virol. 89, 7329–7337 (2015).
Article CAS PubMed PubMed Central Google Scholar
Bao, Y. et al. The influenza virus resource at the National Center for Biotechnology Information. J. Virol. 82, 596–601 (2008).
Article CAS PubMed Google Scholar
Worobey, M., Han, G.-Z. & Rambaut, A. Genesis and pathogenesis of the 1918 pandemic H1N1 influenza Avirus. Proc. Natl. Acad. Sci. 111, 8107–8112 (2014).
Article ADS CAS PubMed PubMed Central Google Scholar
Lindstrom, S. E., Cox, N. J. & Klimov, A. Genetic analysis of human H2N2 and early H3N2 influenza viruses, 1957-1972: Evidence for genetic divergence and multiple reassortment events. Virology 328, 101–119 (2004).
Article CAS PubMed Google Scholar
Wendel, I. et al. The Avian-Origin PB1 Gene Segment Facilitated Replication and Transmissibility of the H3N2/1968 Pandemic Influenza Virus. J. Virol. 89, 4170–4179 (2015).
Article CAS PubMed PubMed Central Google Scholar
Buonagurio, D. A. et al. Evolution of human influenza A viruses over 50 years: rapid, uniform rate of change in NS gene. Science 232, 980–982 (1986).
Article ADS CAS PubMed Google Scholar
Turnbull, M. L. et al. The Role of the B-Allele of the Influenza A Virus Segment 8 in Setting Mammalian Host Range and Pathogenicity. J. Virol. JVI.01205–16, https://doi.org/10.1128/JVI.01205-16 (2016).
Treanor, J. J., Snyder, M. H., London, W. T. & Murphy, B. R. The B allele of the NS gene of avian influenza viruses, but not the A allele, attenuates a human influenza A virus for squirrel monkeys. Virology 171, 1–9 (1989).
Article CAS PubMed Google Scholar
Li, W. & Godzik, A. Cd-hit: A fast program for clustering and comparing large sets of protein or nucleotide sequences. Bioinformatics 22, 1658–1659 (2006).
Article CAS PubMed Google Scholar
Lu, L., Lycett, S. J., Brown, A. J. L. & Leigh Brown, A. J. Reassortment patterns of avian influenza virus internal segments among different subtypes. BMC Evol. Biol. 14, 16 (2014).
Article PubMed PubMed Central Google Scholar
Lycett, S. J. et al. Estimating reassortment rates in co-circulating Eurasian swine influenza viruses. J. Gen. Virol. 93, 2326–2336 (2012).
Article CAS PubMed PubMed Central Google Scholar
Nobusawa, E. et al. Comparison of complete amino acid sequences and receptor-binding properties among 13 serotypes of hemagglutinins of influenza A viruses. Virology 182, 475–485 (1991).
Article CAS PubMed Google Scholar
WHO. A revision of the system of nomenclature for influenza viruses: a WHO memorandum. Bull. World Health Organ. 58, 585–591 (1980).
Google Scholar
Katoh, K. & Standley, D. M. MAFFT multiple sequence alignment software version 7: Improvements in performance and usability. Mol. Biol. Evol. 30, 772–780 (2013).
Article CAS PubMed PubMed Central Google Scholar
Burke, D. F. & Smith, D. J. A Recommended numbering scheme for influenza A HA Subtypes. PLoS One 9, (2014).
Berman, H. M. The Protein Data Bank. Nucleic Acids Res. 28, 235–242 (2000).
Article ADS CAS PubMed PubMed Central Google Scholar
Wang, S., Ma, J. Z., Peng, J. & Xu, J. B. Protein structure alignment beyond spatial proximity. Sci. Rep. 3 (2013).
Cao, Y. et al. Improved side-chain modeling by coupling clash-detection guided iterative search with rotamer relaxation. Bioinformatics 27, 785–790 (2011).
Article CAS PubMed Google Scholar
McDonald, I. K. & Thornton, J. M. Satisfying hydrogen bonding potential in proteins. Journal of molecular biology 238, 777–793 (1994).
Article CAS PubMed Google Scholar
Magnan, C. N. & Baldi, P. SSpro/ACCpro 5: Almost perfect prediction of protein secondary structure and relative solvent accessibility using profiles, machine learning and structural similarity. Bioinformatics 30, 2592–2597 (2014).
Article CAS PubMed PubMed Central Google Scholar
Bloom, J. D., Drummond, D. A., Arnold, F. H. & Wilke, C. O. Structural determinants of the rate of protein evolution in yeast. Mol. Biol. Evol. 23, 1751–1761 (2006).
Article CAS PubMed Google Scholar

Download references

Acknowledgements

This work was supported by: 1. The CAMS Initiative for Innovative Medicine (CAMS-I2M, 2016-I2M-1-005). 2. National Key Plan for Scientific Research and Development of China (2016YFD0500300). 3. The National Natural Science Foundation of China (31470273, 31671371, 31601043). 4. The National Basic Research Program of China (2015CB910501). 5. The Fundamental Research Funds for the Central Universities (2016ZX310195). We thank all members in Jiang’s lab who gave insightful comments to this manuscript.

Author information

Jiejian Luo and Lizong Deng contributed equally to this work.

Authors and Affiliations

Key Laboratory of Protein & Peptide Pharmaceuticals, National Laboratory of Biomacromolecules, Institute of Biophysics, Chinese Academy of Sciences, Beijing, China
Jiejian Luo & Taijiao Jiang
University of the Chinese Academy of Sciences, Beijing, China
Jiejian Luo
Center for Systems Medicine, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, 100005, China
Lizong Deng, Xiao Ding, Lijun Quan, Aiping Wu & Taijiao Jiang
Suzhou Institute of Systems Medicine, Suzhou, Jiangsu, 215123, China
Lizong Deng, Xiao Ding, Lijun Quan, Aiping Wu & Taijiao Jiang

Authors

Jiejian Luo
View author publications
You can also search for this author in PubMed Google Scholar
Lizong Deng
View author publications
You can also search for this author in PubMed Google Scholar
Xiao Ding
View author publications
You can also search for this author in PubMed Google Scholar
Lijun Quan
View author publications
You can also search for this author in PubMed Google Scholar
Aiping Wu
View author publications
You can also search for this author in PubMed Google Scholar
Taijiao Jiang
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

Jiejian Luo, Lizong Deng, Aiping Wu and Taijiao Jiang conceived the study. Jiejian Luo and Lizong Deng did the computational analysis and wrote the paper. Xiao Ding, and Lijun Quan revised and edited the manuscript. All authors read and approved the manuscript.

Corresponding authors

Correspondence to Aiping Wu or Taijiao Jiang.

Ethics declarations

Competing Interests

The authors declare that they have no competing interests.

Additional information

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Supplementary Materials

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Luo, J., Deng, L., Ding, X. et al. Hydrogen Bond Variations of Influenza A Viruses During Adaptation in Human. Sci Rep 7, 14295 (2017). https://doi.org/10.1038/s41598-017-14533-3

Download citation

Received: 11 August 2017
Accepted: 11 October 2017
Published: 30 October 2017
DOI: https://doi.org/10.1038/s41598-017-14533-3

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.